Welcome to the December 9, 2022, edition of ACM TechNews, providing timely information for IT professionals three times a week.
ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version.
The online version now has a button at the top labeled "Show Headlines."
|
|
World's Largest Computing Society Honors 2022 Distinguished Members for Ground-Breaking Achievements, Longstanding Participation
ACM December 7, 2022
ACM has cited 67 longstanding Distinguished Members for contributions that have driven innovation, enhanced computer science (CS) education, and furthered the CS discipline. Said ACM president Yannis Ioannidis, “Each of these new 67 Distinguished Members have been selected for specific and impactful work, as well as their longstanding commitment to being a part of our professional association.” ACM named this year's recipients for advancing algorithms, cybersecurity, data management, energy-efficient computer architecture, data retrieval, healthcare information technology, knowledge graph and semantic analysis, mobile computing, and software engineering. Distinguished Members are expected to have served as mentors and role models, and to have contributed to the field outside the norm.
|
What AI-Generated COVID News Tells Us That Journalists Don't
McGill University (Canada) December 6, 2022
An artificial intelligence (AI) developed by researchers at Canada's McGill University identified biases in news reporting on COVID-19. The AI created simulated news coverage using headlines from CBC articles as prompts. In comparing the simulated news stories to actual CBC reporting, the researchers found that CBC journalists focused more on personalities and geo-politics, while the AI produced more disease-centered reporting. McGill's Andrew Piper said, "Reporting on real-world events requires complex choices, including decisions about which events and players take center stage. By comparing what was reported with what could have been reported, our study provides perspective on the editorial choices made by news agencies."
|
Apple Details Plans to Beef Up Encryption of Data in iCloud
The New York Times Tripp Mickle December 7, 2022
Apple will expand its end-to-end encryption system in order to render most iCloud data unreadable, even when stored in datacenters. Although Apple had not fully encrypted the data so customers can more easily retrieve information for users who were locked out or lost account access, escalating breaches and more data migrating to the cloud prompted the company to fortify its security. The optional Advanced Data Protection framework was designed to shield data of public figures who hackers may target. The upgrade could potentially conflict with the U.S. government and other regimes. Apple has refused to help law enforcement unlock iPhones, while meeting many requests for iCloud backups that include unencrypted messages and photos.
*May Require Paid Registration
|
Soft Robot Detects Damage, Heals Itself
Cornell Chronicle David Nutt December 7, 2022
A soft robot developed by Cornell University engineers can detect when and where damage to its body occurs and immediately heal itself. The SHeaLDS (self-healing light guides for dynamic sensing) robot is comprised of fiber-optic sensors that can identify material deformations, as well as a polyurethane urea elastomer with hydrogen bonds for rapid healing and disulfide exchanges for strength. In tests in which the robot's leg was punctured six times, each cut was detected and self-healed within approximately one minute. In addition, the robot adapted its gait based on the damage without intervention. Cornell's Rob Shepherd said machine learning algorithms able to identify tactile events eventually will be integrated with SHeaLDS to produce "a very enduring robot that has a self-healing skin,” which it also uses “to feel its environment to be able to do more tasks."
|
Coupled Computer Modeling Predicts Coastal Flooding More Accurately
Louisiana State University December 5, 2022
A coupled computer modeling approach used by Louisiana State University (LSU) researchers to recreate coastal flooding caused by 2018's Hurricane Florence was found to be more accurate than traditional modeling techniques. The Category 1 storm caused a compound flooding event involving the convergence of rainfall-swollen rivers and the rising hurricane storm surge along North Carolina's Cape Fear River Basin. The coupled modeling approach runs river and ocean simulations simultaneously, with each simulation providing feedback to the other. The researchers found that the coupled model was 20% to 40% more accurate than traditional models in simulating water levels at the head of the Cape Fear Estuary during Hurricane Florence.
|
Cosmic Rays Are Quantum Computers' Kryptonite
IEEE Spectrum Charles Q. Choi December 7, 2022
University of Chicago researchers have developed a method to reduce the rate of catastrophic quantum computing errors attributable to cosmic rays from about once every 10 seconds to once every 51 days. The strategy involves dividing quantum computers into multiple data chips, each containing multiple superconducting qubits. To monitor their performance, the data chips are linked to an "ancilla chip" that contains additional superconducting qubits. The data chips use conventional quantum error correction codes to handle regular errors, with an additional quantum error correction code to provide protection from cosmic rays. Following a cosmic ray event, the ancilla chip and the data chips not disrupted by the strike collaborate to correct affected data chips and restore the quantum computer's data.
|
Towards Automatic Detection of Road Features with Deep Learning
Hosei University (Japan) December 7, 2022
Researchers in Japan generated a deep learning model that uses high-precision three-dimensional maps to detect road features in point clouds as training data. The researchers used the CloudCompare software to separate the ground surface from point cloud data, then produced area data from the map and extracted component points of features labeled as road signs, traffic lights, or other features. They generated training data by extending area data corresponding to component points, then further assembled point cloud projection images and built the identification model with a YOLOv3 object-detection algorithm. Hosei University's Ryuichi Imai said, "The feature extraction can be done automatically, including the features at undeveloped road map sections."
|
Secret to STEM Diversity May Lie in Peer Mentorship
University of Massachusetts Amherst December 6, 2022
University of Massachusetts, Amherst (UMass Amherst) researchers found that peer mentorship positively affects female science, technology, engineering, and math (STEM) students from their undergraduate through postgraduate lives. The researchers paired 150 female engineering majors with 58 student mentors (32 women and 26 men) and tracked them for eight years. Female mentees with female mentors reported a much greater sense of camaraderie, motivation, and confidence after the end of their first year. They also obtained professional internships and were more likely to complete an undergraduate degree in a STEM field. "From high-quality peer relationships within the academic environment, especially relationships with peers who share a common identity, comes the confidence and motivation to persist, which lasts for a very long time, powering that student through her academic and early professional career," concluded UMass Amherst's Nilanjana Dasgupta.
|
LiDAR Technique Could Help Robotic Vehicles Land Safely on Mars
Optica December 6, 2022
Researchers at the NASA Langley Research Center have developed a LiDAR technique to improve the safety of robotic vehicle landings during future Mars or lunar missions. Typically, flash LiDAR requires a large mechanical gimbal to scan the target landing area and stitch together individual LiDAR images to identify hazardous terrain features when choosing the safest landing location. To eliminate the need for a scanning gimbal and prevent motion blur, the researchers developed a super-resolution algorithm that broadens the LiDAR field of view, takes images of the same scene from various positions and angles, and computationally combines the images to generate a high-resolution view. The algorithm's image resolution was found to be 25 times greater than flash LiDAR alone.
|
Large Language Models Help Decipher Clinical Notes
MIT News Rachel Gordon December 1, 2022
Massachusetts Institute of Technology (MIT) researchers tapped large language models to disentangle unstructured clinical notes in electronic health records, in order to extract meaningful information. The researchers employed a GPT-3-style model to execute tasks such as expanding overloaded jargon and acronyms and extracting drug regimens. Said MIT's David Sontag, "The research team's advances in zero-shot clinical information extraction makes scaling possible. Even if you have hundreds of different use cases, no problem—you can build each model with a few minutes of work, versus having to label a ton of data for that particular task." MIT's Hunter Lang said the researchers also devised a method of formatting task prompts so the model’s outputs are in the proper format.
|
Researchers Develop Scaled-up Spintronic Probabilistic Computer
Tohoku University (Japan) December 7, 2022
Scientists at Japan's Tohoku University, Italy's University of Messina, and the University of California, Santa Barbara have engineered a scaled-up probabilistic computer (p-computer) with stochastic spintronic devices. The researchers demonstrated how stochastic magnetic tunnel junction (sMTJ)-based probabilistic bits (p-bits) can be integrated with field-programmable gate arrays (FPGAs) to implement larger p-bit networks in hardware. They also executed a simulated quantum annealing algorithm in heterogeneous MTJ + FPGA p-computers with systematic assessments for hard combinatorial optimization problems. The researchers benchmarked sMTJ-based p-computer performance against that of classical computing hardware, including graphics processing units and Tensor Processing Units, and found it yields superior throughput and power consumption compared to conventional technologies.
|
Endangered Whales Get Lifeline from High-Tech Lobster Traps
The Wall Street Journal Eric Niiler December 8, 2022
Thirty lobstermen are actively testing ropeless fishing equipment engineered to protect the endangered North Atlantic right whale, in a program overseen by the U.S. National Oceanic and Atmospheric Administration. The lobster traps are designed to prevent entanglement by keeping buoys and ropes stowed underwater until it is time to check the traps. Meanwhile, researchers at the Woods Hole Oceanographic Institution are developing technologies that could shield whales from collisions with ships. One team built a shipboard thermal-imaging system that detects whales by spout heat and alerts ship captains to their presence, while another group deployed a network of acoustic-sensor buoys and underwater drones along the East Coast to listen for whale sounds and transmit the whales' locations to local vessels.
*May Require Paid Registration
|
Go-Based Botnet Exploiting Dozens of IoT Vulnerabilities to Expand Network
The Hacker News Ravie Lakshmanan December 7, 2022
Researchers at Fortinet FortiGuard Labs identified a Go-based botnet that is taking advantage of 21 security vulnerabilities in Internet of Things devices and other software, singling out Windows and Linux operating systems in its efforts to assume control of the affected devices. Fortinet's Cara Lin said the Zerobot botnet "contains several modules, including self-replication, attacks for different protocols, and self-propagation. It also communicates with its command-and-control server using the WebSocket protocol." The vulnerabilities affect a range of devices, including TOTOLINK routers, Zyxel firewalls, F5 BIG-IP, Hikvision cameras, FLIR AX8 thermal imaging cameras, D-Link DNS-320 NAS, and Spring Framework. Added Lin, "Within a very short time, it was updated with string obfuscation, a copy file module, and a propagation exploit module that make[s] it harder to detect and gives it a higher capability to infect more devices."
|
|